Render-Reachable WebAssembly Cyclomatic Complexity

render-cyclomatic-complexity.wasm reports the static branch complexity of the local Wasm code that a module's exported render function can call. It reads a WebAssembly module and emits one unsigned decimal integer, with no trailing newline in the output file.

make -j qip components/application/wasm/render-cyclomatic-complexity.wasm

qip run -i components/text/e164.wasm \
  -o complexity.txt \
  -- components/application/wasm/render-cyclomatic-complexity.wasm

cat complexity.txt

4

The caller decides what value is acceptable. For example, a policy that requires complexity below 22 rejects an output of 22 or more.

How It Calculates The Value #

The reporter starts at the defined exported render function and follows each direct local call and return_call. It visits each reachable function body once, so splitting one function into helpers does not reduce the value.

The value starts at one. It then adds:

• one for each if instruction;
• one for each br_if instruction; and
• one fewer than the number of distinct branch destinations in each br_table.

br, loop, and direct calls do not add branch complexity themselves. A loop contributes through the decision that exits or repeats it. Recursive calls are safe to analyze: the reporter counts the reachable function body once rather than expanding the cycle forever.

The result describes local Wasm instructions only. A direct call to an imported function has no local body to inspect, so it contributes no branch sites. The reporter traps on reachable call_indirect, return_call_indirect, or call_ref; their possible local targets cannot be established from direct-call edges alone. Use the strict profile when those instructions must be prohibited for the complete module.

The input must contain a defined function export named render. Malformed Wasm or a module without that export traps. Validate untrusted bytes with wasm-validate-core-1.0.wasm first.

When Not To Use It #

This number is a static control-flow budget. It does not measure executed instructions, loop iteration counts, memory use, test coverage, or algorithmic cost. Use wasm-counts.wasm for a broad static inventory, wasm-bounded-loops.wasm for recognized loop bounds, and runtime benchmarks for observed cost.